| Author |
Message |
senorverde
Joined: Dec 10, 2011
Posts: 3
Location: NY
|
 Posted: Sat Dec 10, 2011 6:05 am Post subject:
Can It Be Done? Single Supply Op-Amp Filter |
  |
|
Hello! I'm new here and am in need of some help for my project. Lately, I have been considering adding a LP filter to my DYI synth and have been doing some research on all the various types. After giving it some thought, I've decided on using a Sallen-Key filter, although I'm a bit hesitant about how the op-amps have a dual power supply. So, doing more research (the more the merrier  ), I've found that there are these single supply op amps out there that require far less power than their dual power relatives.
My question is how, in the long run, will using single supply op-amps differ from using dual supply op amps? I have heard that there's a lot more distortion, but I am new with these magical devices.
Any suggestions?
Thanks. |
|
|
Back to top
|
|
 |
MadScientist
Joined: Nov 28, 2011
Posts: 63
Location: Denmark
|
| Posted: Sun Dec 11, 2011 12:37 am Post subject:
Re: Can It Be Done? Single Supply Op-Amp Filter |
|
|
| senorverde wrote: | | My question is how, in the long run, will using single supply op-amps differ from using dual supply op amps? I have heard that there's a lot more distortion, but I am new with these magical devices. |
The quick answer is that in the long run you will be shooting yourself in the foot by not immediately going for a dual polarity supply. The only drawback of doing so is that you need a slightly more complicated PSU now.
A Sallen-Key filter can certainly be built using a single supply, and I will give a schematic if you wish. But I think you should hear the trade-offs first.
At the end of the day there is only *one* major difference between dual and single supply op-amps. That difference is how close the inputs and outputs can get to either supply line before bad things happen. An op-amp does actually not care in the least whether you use a single or a dual supply, it only cares about the voltages at the supply rails and inputs/outputs.
So for instance a standard TL072/074 has hard limits to the output voltage. It can only get within about 2V below the positive rail, and 2V above the negative. So if you powered a TL071 from +/-12V, then the output could only go between roughly +/-10V. If on the other hand you ran the same TL071 from a 9V battery, single supply, then the output voltage is limited to between +2V (low) and +7V (high). The signal would 'flat top', meaning heavy distortion, whenever the op-amp tried to drive the signal above or below these limits.
The 'single supply' op-amps simply have the ability to go pretty much all the way toward the supply rails, so from 0 o 9V on a 9V battery. The price they pay is that to get those last 4V of swing, 2 at either supply rail, they need to use special techniques, which increases distortion for signals above and below the 2V 'rail limits'. However, if you insisted on running an op-amp at, say, 5V, then you pretty much wouldn't have a choice except using a 'single supply' op-amp. A TL071 would in this case only have about 1V of voltage travel on the output, as the 2V limits are absolutes, and not a percentage.
So why don't we run our stuff from a single supply all the time? As it happens, you *could* build, say, a nice synth on a 0-20V supply, with signals centered around 10V instead of ground. However, if you did so, you would quickly discover a funny fact. In order to keep all signals centered on some arbitrary voltage, here 10V, your circuits would quickly become way more complicated to the point, where it is much easier to just get a dual supply.
Because that is pretty much what a dual supply is, an artificial mid point right in the center between the supply rails, and where you can draw different currents from each rail (which is highly convenient in many situations). If you had a good reason, then you could also build a synth running on +20/+32/+44V, which as far as the op-amps are concerned would change absolutely nothing compared to using plain old -12/0/+12V. The voltage differences would be the same. So would be using 0/+12/+24V, which would be a 'single supply' with an artificial mid-point at 12V.
So you'd use single-supply op-amps if you are artificially constrained with the voltages available, say by battery voltage in portable equipment, and pay the price in added circuit complexity.
For everyday operation you'd use normal op-amps with either single or dual supplies, and modify the schematics accordingly. Personally I'd set the bar for using dual supplies very low. I'd prefer to using, say, two 9V batteries, instead of trying to modify a schematic and having to use special op-amps in order to run stuff from a single battery. There are other problems with single supply setups, but that would require some schematic examples to demonstrate.
Hope this makes some kind of sense.
Frank. |
|
|
Back to top
|
|
|
bod
Joined: Apr 28, 2009
Posts: 148
Location: Glasgow
|
| Posted: Mon Dec 12, 2011 4:47 pm Post subject:
|
|
|
Firstly, does it have to be single supply? Is that what the rest of your synth is based on?
Secondly, and despite mad scientists insightful explanation (which is a great read, I didn't quite realise that before!!) I've had great result with running a filter based around the lm13700 on a single supply. With a few extra caps/resistors for a basic filter than a sallen key, you've got a voltage controlled filter and no need for the dual gang pot. Sometimes the distortion can be a feature as well.
Alternatively there is ray wilsons filter from the WSG, and the ms20 style filter from the monotron that should be of interest.
Happy filtering!! |
|
|
Back to top
|
|
|
senorverde
Joined: Dec 10, 2011
Posts: 3
Location: NY
|
| Posted: Tue Dec 13, 2011 3:53 am Post subject:
|
|
|
| bod wrote: | Firstly, does it have to be single supply? Is that what the rest of your synth is based on?
Secondly, and despite mad scientists insightful explanation (which is a great read, I didn't quite realise that before!!) I've had great result with running a filter based around the lm13700 on a single supply. With a few extra caps/resistors for a basic filter than a sallen key, you've got a voltage controlled filter and no need for the dual gang pot. Sometimes the distortion can be a feature as well.
Alternatively there is ray wilsons filter from the WSG, and the ms20 style filter from the monotron that should be of interest.
Happy filtering!! |
Wow, first off thank you for everyone's answers!
The reason why it has to be a single op amp supply is because of the limited voltage of the entire circuit. You see, I'm in a little wager with a few friends to see which one can come up with the best sounding 'synth' using the the smallest amount of components. For this reason, I've decided to go with a 74 series of inverters and a Sallen Key filter. Now, as we all know the 74 series don't require that much power for them to make a noise. Op amps however, especially the dual supply kind, require supply voltages three or more times greater than the VCO itself! Their single supply cousins however do not, with supply voltages in the same neighbourhood as the VCO itself. So, in the end it'll probably be a choice between the following:
>Single, low power supply with a lot of distortion.
OR
>Dual supply, considerable power supply with virtually no distortion. |
|
|
Back to top
|
|
|
MadScientist
Joined: Nov 28, 2011
Posts: 63
Location: Denmark
|
| Posted: Tue Dec 13, 2011 4:58 am Post subject:
|
|
|
Are you allowed to use a micro-controller?  Some of them come with integrated D/A converter, which means you'd only need a few capacitors plus the MCU itself for a fully fledged synthesizer.
If MCUs aren't allowed, then how do you count op-amps or logic-ICs? Each one hides varying levels of integration. A fully discrete component synth design, running from a single power rail, would probably win the day if you count *all* the components, including those inside integrated circuits.
Hmm...?  |
|
|
Back to top
|
|
|
senorverde
Joined: Dec 10, 2011
Posts: 3
Location: NY
|
| Posted: Tue Dec 13, 2011 6:22 am Post subject:
|
|
|
| MadScientist wrote: | Are you allowed to use a micro-controller? Some of them come with integrated D/A converter, which means you'd only need a few capacitors plus the MCU itself for a fully fledged synthesizer.
If MCUs aren't allowed, then how do you count op-amps or logic-ICs? Each one hides varying levels of integration. A fully discrete component synth design, running from a single power rail, would probably win the day if you count *all* the components, including those inside integrated circuits.
Hmm...? |
Now that's a really good idea! However, they are not allowed (drat!). Basically, the vein of the wager is to take a bunch of ICs (inverters, gates, &c.) coupled with components and make them sound halfway decent through whatever means you see fit. I'll look into the details soon. |
|
|
Back to top
|
|
|
Tim Servo
Joined: Jul 16, 2006
Posts: 924
Location: Silicon Valley
Audio files: 11
|
| Posted: Tue Dec 13, 2011 1:20 pm Post subject:
Can It Be Done? Single Supply Op-Amp Filter |
|
|
You really should check out the Lunetta sub-forum here if you haven't already. Probably find a TON of great logic-based, single-supply noisemakers.
Also, someone posted a VERY simple schematic showing a VCF that was basically a Vactrol, a resistor, and a cap. I'll look around and see if I can find it (I think the thread was "Simplest Filter" or something like that...). It sounded surprisingly good and was simple as dirt. Even if you made your own Vactrol from a photocell and an LED, you could implement a VCF and a VCA very easily this way. Also, look into using the HF11F1 photocoupler. This is a 'vactrol' in an easy to use DIP package, but it uses a phototransistor instead of a photocell has much quicker response time.
http://search.digikey.com/us/en/products/H11F1M/H11F1M-ND/1793945
I would think something like that would work nicely. 
Tim (also very simple) Servo |
|
|
Back to top
|
|
|
Tim Servo
Joined: Jul 16, 2006
Posts: 924
Location: Silicon Valley
Audio files: 11
|
|
|
Back to top
|
|
|
comrade_zero
Joined: Mar 05, 2009
Posts: 66
Location: arizona
Audio files: 4
|
| Posted: Tue Dec 13, 2011 7:22 pm Post subject:
have you thought of a cd4069? |
|
|
Have you considered using a cd4069?
This is not an op-amp solution, but it runs off of a low-voltage single supply and can make a really good resonant filter. Extremely low parts count.
Simple to build and easy to use.
The most common example is the Wasp filter, in which ca3080's are used as variable resistors in a topolgy that mimics a two-pole state variable filter. We don't need all that do we?
If we replace the 3080's with a dual ganged pot we can control the cutoff just fine. Or we use LED/LDR's for poor man's voltage control...
If you are interested check around, you will find a couple schematics, or I could probably find time to post one soon...
My current version is 1 IC, two caps, two resistors, and 2 pots, for a two-pole filter with Lowpass and Bandpass. I can run it as low as +5 v on a single supply...
cheers
c_z |
|
|
Back to top
|
|
|
Mongo1
Joined: Aug 11, 2011
Posts: 411
Location: Raleigh NC
|
| Posted: Wed Dec 14, 2011 8:24 am Post subject:
|
|
|
You know, after hanging out on this forum for a while, I'm becoming more and more convinced that our most prolific contributors are based in a facility for the criminally insane.
I love it!
BTW - the vactrol idea is awesome. I think that is far and away the best approach for this problem. The sound quality to parts count ratio will beat the daylights out of your competition, and most of them will never think of that approach.
Good luck.
Gary |
|
|
Back to top
|
|
|
JingleJoe
Joined: Nov 10, 2011
Posts: 878
Location: Lancashire, England
Audio files: 14
|
|
|
Back to top
|
|
|
MadScientist
Joined: Nov 28, 2011
Posts: 63
Location: Denmark
|
| Posted: Wed Dec 14, 2011 12:59 pm Post subject:
|
|
|
| There is a component called an analog comparator, which is better at comparing than any op-amp is likely to be. One example would be the LM393. |
|
|
Back to top
|
|
|
inlifeindeath
Joined: Apr 02, 2010
Posts: 316
Location: Albuquerque, NM
|
|
|
Back to top
|
|
|
Mongo1
Joined: Aug 11, 2011
Posts: 411
Location: Raleigh NC
|
| Posted: Wed Dec 14, 2011 1:20 pm Post subject:
|
|
|
Hi Joe -
If I understand what you're saying (and who knows if that's true?) it sounds like you're trying to use the sluggish nature of the vactrol as a filter. That seems like it would work, but the corner frequency of the filter would be fixed.
I don't know how useful that would be.
I was assuming the idea was more in the nature of using the vactrol to replace the R part of normal filter.
Gary (hang on - the nurse is here with my meds) |
|
|
Back to top
|
|
|
JingleJoe
Joined: Nov 10, 2011
Posts: 878
Location: Lancashire, England
Audio files: 14
|
| Posted: Wed Dec 14, 2011 2:56 pm Post subject:
|
|
|
| Mongo1 wrote: | Hi Joe -
If I understand what you're saying (and who knows if that's true?) it sounds like you're trying to use the sluggish nature of the vactrol as a filter.
|
no no no no  the LDR part of the vactrol would control a filter, I was merely saying that the "slugish nature" stops the PWM signal form modulating the filter and should in theory let the light level* control the filter.
(*which is controlled by the PWM signal, which will have to be around 100 to 500 Hz for good operation I think)
| Mongo1 wrote: |
I was assuming the idea was more in the nature of using the vactrol to replace the R part of normal filter.
Gary (hang on - the nurse is here with my meds) |
Thats exactly what I ment |
|
|
Back to top
|
|
|
JingleJoe
Joined: Nov 10, 2011
Posts: 878
Location: Lancashire, England
Audio files: 14
|
| Posted: Wed Dec 14, 2011 3:01 pm Post subject:
|
|
|
| MadScientist wrote: | | There is a component called an analog comparator, which is better at comparing than any op-amp is likely to be. One example would be the LM393. |
Brilliant thats just the ticket!
Yes I am certain others have done this before despite this I could find no explanation as clear as my graph/diagram/description around the internets. |
|
|
Back to top
|
|
|
Tim Servo
Joined: Jul 16, 2006
Posts: 924
Location: Silicon Valley
Audio files: 11
|
| Posted: Wed Dec 14, 2011 6:18 pm Post subject:
Can It Be Done? Single Supply Op-Amp Filter |
|
|
Not quite sure I understand the idea of using PWM to control the LED in a vactrol. Why not just directly control the brightness of the LED with the DC control voltage? PWM control has advantages in other areas (like motor control or high power LEDs), but I think it's an unnecessary step here. Right? Wrong? I have seen circuits that essentially switch a resistor in and out with a varying frequency and/or pulse width, allowing voltage control to replace a standard pot, but that's not the case here. Also, I think with the average vactrol, the response is so slow that a pulse frequency of anything more than a few Hertz would be lost (would probably work much better with the ultra fast H11 optocoupler).
Or am I just typically confused?
Tim (confused, befuddled and bewildered am I) Servo |
|
|
Back to top
|
|
|
Estebandito
Joined: Dec 25, 2017
Posts: 33
Location: Amsterdam
Audio files: 3
|
|
|
Back to top
|
|
|
|
Forum index » DIY Hardware and Software
FAQ
Calendar
Search
Memberlist
Usergroups
Register
Profile
Log in to check your private messages
Log in
Chat Room
